mirror of
https://github.com/ytdl-org/youtube-dl
synced 2024-11-05 09:57:59 +09:00
96 lines
3.3 KiB
Python
96 lines
3.3 KiB
Python
# coding: utf-8
|
|
from __future__ import unicode_literals
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import (
|
|
int_or_none,
|
|
merge_dicts,
|
|
mimetype2ext,
|
|
url_or_none,
|
|
)
|
|
|
|
|
|
class AparatIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?:www\.)?aparat\.com/(?:v/|video/video/embed/videohash/)(?P<id>[a-zA-Z0-9]+)'
|
|
|
|
_TESTS = [{
|
|
'url': 'http://www.aparat.com/v/wP8On',
|
|
'md5': '131aca2e14fe7c4dcb3c4877ba300c89',
|
|
'info_dict': {
|
|
'id': 'wP8On',
|
|
'ext': 'mp4',
|
|
'title': 'تیم گلکسی 11 - زومیت',
|
|
'description': 'md5:096bdabcdcc4569f2b8a5e903a3b3028',
|
|
'duration': 231,
|
|
'timestamp': 1387394859,
|
|
'upload_date': '20131218',
|
|
'view_count': int,
|
|
},
|
|
}, {
|
|
# multiple formats
|
|
'url': 'https://www.aparat.com/v/8dflw/',
|
|
'only_matching': True,
|
|
}]
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
|
|
# Provides more metadata
|
|
webpage = self._download_webpage(url, video_id, fatal=False)
|
|
|
|
if not webpage:
|
|
# Note: There is an easier-to-parse configuration at
|
|
# http://www.aparat.com/video/video/config/videohash/%video_id
|
|
# but the URL in there does not work
|
|
webpage = self._download_webpage(
|
|
'http://www.aparat.com/video/video/embed/vt/frame/showvideo/yes/videohash/' + video_id,
|
|
video_id)
|
|
|
|
options = self._parse_json(
|
|
self._search_regex(
|
|
r'options\s*=\s*JSON\.parse\(\s*(["\'])(?P<value>(?:(?!\1).)+)\1\s*\)',
|
|
webpage, 'options', group='value'),
|
|
video_id)
|
|
|
|
player = options['plugins']['sabaPlayerPlugin']
|
|
|
|
formats = []
|
|
for sources in player['multiSRC']:
|
|
for item in sources:
|
|
if not isinstance(item, dict):
|
|
continue
|
|
file_url = url_or_none(item.get('src'))
|
|
if not file_url:
|
|
continue
|
|
item_type = item.get('type')
|
|
if item_type == 'application/vnd.apple.mpegurl':
|
|
formats.extend(self._extract_m3u8_formats(
|
|
file_url, video_id, 'mp4',
|
|
entry_protocol='m3u8_native', m3u8_id='hls',
|
|
fatal=False))
|
|
else:
|
|
ext = mimetype2ext(item.get('type'))
|
|
label = item.get('label')
|
|
formats.append({
|
|
'url': file_url,
|
|
'ext': ext,
|
|
'format_id': 'http-%s' % (label or ext),
|
|
'height': int_or_none(self._search_regex(
|
|
r'(\d+)[pP]', label or '', 'height',
|
|
default=None)),
|
|
})
|
|
self._sort_formats(
|
|
formats, field_preference=('height', 'width', 'tbr', 'format_id'))
|
|
|
|
info = self._search_json_ld(webpage, video_id, default={})
|
|
|
|
if not info.get('title'):
|
|
info['title'] = player['title']
|
|
|
|
return merge_dicts(info, {
|
|
'id': video_id,
|
|
'thumbnail': url_or_none(options.get('poster')),
|
|
'duration': int_or_none(player.get('duration')),
|
|
'formats': formats,
|
|
})
|